The overall goals of this proposal are to determine if an oligonucleotide-based microarray can be used for the discovery of in vivo genomic transcription factor binding sites and, if so, to use the arrays to identify all genomic binding sites for specific human transcription factors. In brief, our experiments are based on using chromatin immunoprecipitation to selectively enrich for all the binding sites in the human genome of a particular transcription factor. After immunoprecipitation, the fragments will be labeled and used to probe a genomic microarray (i.e. a ChiP-chip assay). Positive signals will identify genomic regions that contain the binding sites. Sequence comparisons of the identified genomic regions will allow the development of a consensus binding site. Also, our studies can provide information as to which binding sites for different factors are commonly clustered on a genome-wide basis. [unreadable] [unreadable] Our proposal will be divided into three phases; proof of concept, development of a first exon identification method, and discovery of functional elements. The size of the human genome essentially precludes the use of spotted PCR fragments for the development of comprehensive promoter-specific human microarrays. Clearly, the development of high density oligonucleotide arrays are essential if one wishes to perform a comprehensive identification of in vivo binding sites for specific human transcription factors. Therefore, our first Aim focuses on determining that oligonucleotide arrays created using the NimbleGen Maskless Array Synthesis technology can be used in ChiP-chip assays. Our second Aim is focused on the development of a first exon identification method using PromotedExon Discovery Arrays that span the entire ENCODE-selected sequence. The identification of all the first exons utilized in a particular cell type will greatly aid in the interpretation of the ChiP-chip data obtained in Aim 3. Finally, in Aim 3 we propose to use the Promoter-Specific Arrays to globally identify binding sites for numerous human transcription factors. We will then use this information to identify functional sequence elements and common promoter architectures. [unreadable] [unreadable] [unreadable]